Given the importance of maintaining the item difficulties of the English Section of College Scholastic Ability Test (hereafter CSAT) even after it was changed to criterionreferenced assessment for the 2018 academic year, this study aims to analyze ...
Given the importance of maintaining the item difficulties of the English Section of College Scholastic Ability Test (hereafter CSAT) even after it was changed to criterionreferenced assessment for the 2018 academic year, this study aims to analyze the distribution of words used in the English section of the 2015-2017 CSATs and to see whether there is a significant association between use of graded words and the three tests.
To this end, the basic words in the 2009 Revised National English Curriculum are graded on the basis of Nation’s (2012) BNC/COCA headword lists into two or four groups. The vocabulary analysis program, Range32H, is used to investigate the frequency of the graded word types and tokens.
Chi-square tests showed the following results. First, there is no association between the frequency of the graded word types and tokens to the English section of the 2015-2017 CSATs. Second, even though not statistically significant, the distribution of graded words in the reading passages was more consistent than that in the listening scripts. Finally, the TTRs of the listening scripts and three reading passages in the three tests are almost the same. This implies that lexical diversity in the three tests is quite ideal.
The findings suggested that if other variables that affect the item difficulty such as the syntactic complexity of the sentence and the average number of sentences in a passage are wellcontrolled, it will contribute to test equating in the English section of the CSAT. The findings also suggested that linguistic analysis for test equating should be used as a tool for estimating the level of item difficulty for the criterion-referenced assessment of English in the CSAT.